THE CONSTITUTION OF THE COPPEE-TIN SERIES OF ALLOYS. 
51 
of the alloy ; they oxidise more slowly than the ground, so that they probably contain 
more tin. They are also blackened by FeClg, a feature that becomes very marked in 
the succeeding alloys. 
Sn 21. S.c. chill at 725° (fig. 54). 
This ingot was very much distorted by granulation in the chill, a proof that it was 
far from solid at that moment. Polishing etches out the surface into a pattern of 
primaries, leaving these faintly cop23er-coloured on a dead white ground, and heat- 
oxidation darkens the primaries more rapidly than the ground, a proof that they are 
richer in copper. FeClg is rather variable in its action on them, but, on the wliole, it 
darkens them as compared with the ground. These primaries (fig. 54) consist of 
rounded blobs, often isolated, but sometimes grouped in oblique sprays, unlike the 
rectangular yS. The lobes of primary, when examined with a higher power, are 
uniform, not showing any tendency to break up into the bar pattern of the lower 
chills, but in the chill primary, or in the mother-substance, one can see the beginning 
of the bar pattern.* 
Sn 21. S.c. chill at 710° (not reproduced). 
This is a smooth solid-looking lump, but cutting and polishing develop large 
cracks; polishing reveals no pattern. FeClg acts very slowly, but finally brings out 
polygonal patches and some rotation effect, but no bars or other pattern. The ingot 
is a good example of a uniform solid solution, and shows that the solidus lies above 
710°. The chill at 700° was uniform. That at 650° was a pattern of large patches 
(fig. 55), but each of these patches had a uniform surface, with no bars or other 
pattern. There is, therefore, a region of temperature in which the alloy is a uniform 
solid solution. Probably on account of fragments breaking off’ from the edges of the 
cracks, vm found it impossible to avoid scratches in polishing this ingot. 
Sn 21. S.c. chill at 590° (fig. 56). 
Here we are a little below the transformation curve D'E', and several per cent, of 
the new rj body has crystallised out of tlie solid solution in curved bands which 
coincide with the previous polygonal boundaries, and in groups of parallel bars which 
generally start from a boundary. In the photograph (fig. 56), which is taken from a 
FeClg etch, these are black, but they are richer in tin than the ground, for if the 
surface is heated, these t] bars oxidise more slowly than the ground, remaining white 
while the ground turns a rich brown. The photograph tends to give a false impression 
as to the amount of the new material, as it was taken from a patch of the surface 
where it was abundant; really, there is less than 10 per cent, of it. 
As the chill was made a little above the D' temperature of 580°, we should not have 
* The striation visible in fig. 54 on the lobes of primary is due to imperfect polishing. 
H 2 
